Covering Disruptive Technology Powering Business in The Digital Age

Written by: Martin Dale Bolima, Tech Journalist, AOPG.

USD $10 billion.

That’s NVIDIA CEO Jensen Huang’s estimate of how much the company spent on Artificial Intelligence (AI) research and development—ultimately birthing the supercharged but very pricey Blackwell, NVIDIA’s next-generation graphic processor for AI.

Again, that’s USD $10 billion!

It’s a staggering amount, and it only shows the lengths Big Tech companies will go to secure pole position in this age of AI.

It’s also a reminder that while AI is big business, it’s also very costly. And there’s no telling how these high costs would impact the final tab of companies pursuing smaller-scale AI investments.

At the moment, all signs point to growing AI adoption, with Dataiku forecasting that spending on AI solutions will reach USD $646 million in 2026—and that’s among companies in Southeast Asia alone. It’s not surprising, as the most successful organisations, according to HubSpot Vice President of Engineering Nadia Alramli, will be “those that implement the necessary infrastructure and processes to embrace and evolve with it, empowering them to enjoy the full potential of AI solutions.”

“AI will change the DNA of society,” she added.

But not if society can no longer afford it.

The Rising Costs of AI Implementation

The good news is that as of today, organisations worldwide seem able and willing to invest in AI—something the Dataiku estimate seems to prove. Research from Stanford University also found that global corporate investment in AI topped USD $100 billion in 2020, 2021, and 2022, underscoring companies’ financial capacity to spend on AI-related innovation and their willingness to do it.

Of course, an AI investment on a smaller scale dwarfs the billions of dollars Big Tech companies pour into their own AI investments (about USD $10 billion and USD $12 billion in planned AI investments for Meta and Google, respectively, this year and USD $14 billion already spent by Microsoft in Q1 of 2024). But it is still substantial, with expenses in hardware and software making up part of the final tab and the integration of AI systems into existing infrastructure adding to the total cost. Licensing for AI platforms, like Google Cloud AI, IBM watsonx, or Microsoft Azure AI, can also be expensive, running up the bill even more—especially when the company scales.

Beyond these initial setup costs are the recurring operational expenses of actually running AI. Energy costs, in particular, can be substantial in light of the computational power required to run and train AI. Skilled labour doesn’t come cheap either, with data scientists, machine learning engineers, and other AI specialists—the people equipped with the knowledge and skills to operationalise AI—commanding high salaries due to their expertise. In fact, based on data aggregated by Glassdoor, the average salary for a machine learning engineer in the US tops USD $120,000 per year, with top-tier professionals earning much more.

Daniel Hein, Chief Architect, Asia Pacific & Japan, Informatica

Add those all up and the cost of implementing AI balloons exponentially—perhaps to levels already out of the reach of smaller businesses. But, if it is any consolation, the cost of deploying AI appears to be going down, at least according to Daniel Hein, Chief Architect, Asia Pacific & Japan, at Informatica, who also claims it is actually possible to reduce the operational costs of running AI.

“The sustainability of AI growth depends on an organisation’s size and resources. This is where automation and resource optimisation can help. By automating routine data management tasks like data discovery, classification and quality checks, smaller organisations can significantly reduce the ongoing operational costs associated with AI initiatives,” Hein told Disruptive Tech News (DTN). Additionally, the cost of AI development and deployment is constantly decreasing due to advancements in cloud computing, open-source tools, and pre-trained models. This not only makes AI more accessible but also helps achieve a faster ROI.”

Are Economies of Scale at Play in the Future?

Jess O’Reilly, Area Vice President for Asia, UiPath

Jess O’Reilly, Area Vice President for Asia at UiPath, is even more bullish, anticipating the economies of scale to offset whatever expenses companies would incur from deploying AI moving forward and even lower it to more manageable levels—particularly here in Southeast Asia.

“AI advancements continue to drive efficiencies and create more opportunities, fuelling an innovation flywheel across industries in the region. As technologies combining AI and automation mature and become more mainstream, we anticipate economies of scale and productivity gains that would eventually mitigate the costs associated with AI development and deployment,” O’Reilly told DTN. “A combination of AI and automation can help unlock insights from large data volumes and allow organisations to identify opportunities for effective optimisation across functions. This helps lower organisations’ operational costs, mitigate human errors, and maximise resources, fostering efficiency and driving sustainable growth.”

However, O’Reilly’s optimism on AI reaching the economies of scale stage seems a tad too optimistic. It is possible, yes, but likely for larger enterprises only, who to begin with have the financial capacity and resources to spread the costs of AI across their extensive operations. Amazon, for instance, leverages AI across its vast logistics network, customer service, and product recommendation systems, in turn amortising the high costs of AI over millions of transactions and making the investment both worthwhile and more justifiable.

The same might not apply to Small- to Medium-sized Enterprises (SMEs). These businesses often lack the financial resources to invest heavily in AI, and this limitation generally makes the cost per unit of benefit significantly higher compared to larger companies. Unfortunately, the touted benefits of investing in AI—increased efficiency, automation, and better productivity, among others—might not necessarily bridge this disparity between big enterprises and SMEs.

Left unaddressed, this gap will only widen to the point where only the most financially capable companies can afford to harness the full potential of AI, potentially stifling innovation and competition in the market to the bigger detriment of smaller players.

A Gamble That Can Pay Off Handsomely

For smaller businesses deterred by high AI costs, AI-as-a-Service (AIaaS) offers a viable alternative. These services, like Amazon SageMaker and Azure AI, allow businesses to leverage AI’s potentially game-changing capabilities without the upfront costs of building extensive in-house infrastructure, hiring skilled staff, and more. This, at least in theory, makes advanced AI accessible even to smaller businesses

But, to be clear, utilising AIaaS also comes with its own set of financial considerations. While organisations can reduce upfront costs and ensure scalability, they will still have to account for subscription fees, data storage costs, and potential expenses related to data security and privacy. These expenses may not be as high compared to building AI from scratch, but they can accumulate when not monitored vigilantly. Additionally, since AIaaS solutions are provided by third parties, depending on them for essential business functions can lead to strategic risks, including vendor lock-in, reliance on the provider’s pricing structures, and security gaps.

Matty Kaffeman, Vice President, North Asia and Korea, Verint

So, either way, there are significant costs involved when using AI. Regardless, it might be all worth it, according to Matty Kaffeman, Vice President, North Asia and Korea, at Verint.

“The integration of AI should be viewed as a seamless augmentation to current solutions, technology, and capabilities. By leveraging AI, organisations can enhance customer engagement through personalised interactions, enhance operational efficiency by automating routine tasks, and drive cost reduction by optimising resource allocation,” Kaffeman pointed out to DTN in an exclusive interview.

“This natural evolution empowers businesses to harness the potential of AI to elevate their performance and deliver greater value to customers,” he added. “Embracing AI as a complementary force enables a proactive approach to innovation, setting the stage for continuous improvement and sustainable growth. This also allows businesses to mitigate the risk of ‘pricing itself out’ and achieve a balance between cost and value.”

In other words, AI might just be worth the gamble.

Then again, not all gambles work out that well.

(0)(0)